Subjective everyday experience entails a well-structured and continuous flow of sensory signals, actions, thoughts and emotions. How does the brain build such a coherent representation of space and time despite the vast amount and confusing nature of the input? Furthermore, what are the physiological constraints preventing simultaneous awareness of multiple spatial and temporal instances? Here I propose a novel approach ("brain modes") to investigate these issues within life-like experimental settings. I will investigate how the brain selects and integrates relevant information using complex dynamic environments that includes space, time and multisensorial inputs. Combining model-free and model-driven analyses of functional imaging data I will examine: 1. How signals in different sensory modalities and same/different locations interact in complex environments; 2. How contextual information influences memory encoding and retrieval and the ability to integrate current sensory signals with events in the past. 3. How prospective goals and expectancies arising from the temporal dynamic of the context influence on-line processing. My expectation is that the results will reveal novel mechanisms underlying the ability to organise information in an orderly manner, on a time-line spanning the past, the present and the future; and how we can direct our thoughts along this time-line. My investigation will provide new evidence on the capacity limitations of this selection process and how integration and competition interact to form a representation of the external world that evolves as a coherent flow through space and time. Potential practical implications are foreseen for the design of brain-machine interfaces and for understanding the abnormal perceptions of mental illness.